How PwC Australia Uses AI in Live M&A Engagements

I serve as the Deals CTO in Australia, responsible for the client-facing suite of technology across the Deals business, while remaining a practitioner focused primarily on applying analytics as part of financial and commercial diligence.

A lot of my work is focused on the Technology, Media, and Telecommunications and Healthcare sectors. Day to day that means shaping how our teams apply advanced AI models to digest large, complex sets of information, and using those capabilities as part of diligence work so we can move beyond the legacy analysis of “top ten” customers and toward capability that can deliver contextual understanding of transactional data and documents at scale.

We’ve already been applying analytical techniques for a number of years, which enhanced our ability to deal with numerical data. Harvey provides us with the ability to extend that approach to text, so we can understand not just what happened, but why, and crucially the impact this has on the future of the target business.

AI is redefining materiality thresholds in diligence by making previously cost prohibitive tasks commercially viable. We can now review vast sets of board minutes, internal communications, process documents, and customer agreements to identify both risks and value creation opportunities, not just mitigate downside.

In practice, we use Harvey as both a drop-in replacement for general-purpose prompting with leading foundation models and, more importantly, as our tool of choice for large-scale text extraction and summarisation where quality and volume matter, such as extracting pricing and discount mechanics from SaaS customer contracts to inform revenue trajectory analyses. Harvey’s model-agnostic architecture lets us access the best of OpenAI, Google, and Anthropic without juggling multiple tools, keeping us on the frontier as models evolve rapidly.

Critically, Harvey also provides a secure sandbox with external workspaces and workflow capabilities to bring clients into the process while protecting IP and data, which directly addresses enterprise risk, governance, and collaboration bottlenecks. This technology is helping us shift the industry, and our conversations with clients, from risk mitigation to value creation and realisation. Combining our deep understanding of M&A work with Harvey’s technical expertise has allowed us to build workflows that address key concerns from customers and targets, delivering real value today.

A recent e-commerce due diligence project illustrates the impact well. We used Harvey alongside internal tooling to complete the heavy lift analysis early, which freed us to focus our client discussion on what was actually driving performance. We were able to assess the credibility of a new product line, unpacking pricing and margin dynamics, and interrogating whether declared plans were truly executable. We delivered this insight as part of our work, rather than just spending our time on the analysis without any real value-add conclusions.

Across engagements, we are moving to transcribing diligence calls and vendor/expert discussions and then loading those transcripts and data room documents into Vault workspaces so we can reliably find and cite specific statements, instead of diluting insights because a critical reference can’t be retrieved at the 11th hour. That simple shift to comprehensive ingestion and retrieval has repeatedly transformed “I’m sure I heard X” moments into precise, evidenced answers in minutes.

Three capabilities stand out. First, Harvey’s ability to digest hundreds of thousands of documents and extract structured signals with high fidelity is materially better for large text extraction workflows than generic tools, and this matters on live deals with sprawling data rooms.

Second, secure external workspaces create a governed environment to blend best-in-class models with our proprietary IP, enabling client collaboration without sacrificing confidentiality or competitive advantage.

Third, Workflow Builder allows us to codify practitioner IP into reusable processes, from rapid industry and product research using public support wikis to more advanced PwC IP-guided analyses, so teams and clients can repeatedly run high quality workflows with consistency. Most importantly, we are doing this centrally and globally, as well as at the coal face of client delivery.

We see two compounding effects: scale and depth. On scale, Harvey unlocks workflows that were previously uneconomic. We can run document by document term extraction and expand our scope without expanding headcount, which we expect will move us beyond risk-only diligence to integrated value identification across domains: financial, commercial, tax, HR, and ESG. On depth, faster ingestion and retrieval shift our time from drudgery to judgment, so client conversations focus on implications, not process. This translates to sharper investment theses and faster bid-readiness.

Our clients can benefit twice: They can vet more opportunities earlier, running a “60–70%” version quickly and efficiently and then bringing us in to deliver the “150%” deep dive when conviction and competition require it. That model increases pipeline throughput for clients while improving certainty on the assets that go to the next stage of their investment approval process. And for us, it means delivering more value at the same price by reallocating effort from manual processing to insight generation.

The practical, lived impact for teams is often felt at crunch moments, such as late-night updates or weekend sprints, where Harvey enables rapid, accurate responses and preserves human attention for truly commercial questions, which improves client responsiveness and team sustainability at once.

Aim for frontier models and operating platforms that are model-agnostic. In this market, being even three months behind is too slow, and you need the flexibility to switch or blend engines as they leapfrog each other.

Treat AI as leverage to do more and better work, delivering value for money rather than cost reduction. Target the tasks you historically skipped due to budget or bandwidth constraints, like swapping sample analysis for more comprehensive testing or large scale contract term extraction.

Prioritise secure, governed collaboration so you can expose your IP through workflows and external workspaces. That is how you scale expertise to clients safely and repeatedly without eroding differentiation. It is really important that we treat our clients’ and their investment targets’ data with respect and only use it for appropriate purposes and with the right secure AI models. Harvey provides this out of the box with a permissioned project Vault and secure AI models from the leading providers.

Finally, invest in codifying practitioner and institutional IP into workflows and dedicate high-performing practitioners to embedding day to day methodologies into the platform. Usage typically plateaus after initial access unless you deliberately consider how your teams work.